|
Brian Yamauchi Lead Roboticist iRobot Corporation |
|
The Stingray Project
I'm the Principal Investigator (PI) for
Stingray, a Phase II SBIR project funded by the US Army Tank-Automotive Research, Development, and
Engineering Center (TARDEC). The goal of this research project is to
develop techniques for high-speed teleoperation of small unmanned ground
vehicles (UGVs).
Thousands of small UGVs, like the iRobot PackBot, have been deployed overseas,
where they are helping soldiers deal with improvised explosive devices (IEDs)
and other threats. However, in order to be useful in a wider range of
missions, such as high-tempo infantry operations, small UGVs will need to
operate at much higher speeds.
For the Stingray Project, we are taking a prototype of iRobot's next-generation
Warrior x700 UGV, and
modifying it for high speed operation. In order to control small UGVs at
high speeds, new teleoperation techniques are required. Stingray will
combine immersive teleoperation with semi-autonomous driver assist behaviors.
We are working with Chatten
Associates to integrate their Head-Aimed Remote Viewer (HARV) to the
Stingray Warrior UGV. The HARV combines a head-tracking system with a
head-mounted display and a remote gimbaled camera. The camera tracks
every motion of the operator's head, providing the illusion of being in the
vehicle, and greatly increasing situational awareness.
In addition, we will add semi-autonomous behaviors that will help the operator
control the Stingray Warrior at high speeds. These behaviors will allow
the operator to issue high-level directives to the UGV, such as "maintain
this heading" or "follow this street", freeing the operator from
needing to control every move the vehicle makes. This will allow the
operator to search the environment for potential threats, while the UGV autonomously
controls low-level vehicle motions. These behaviors will also allow
the UGV to maintain a stable speed and heading, even when driving at high speed
over rough terrain.
This video shows our initial experiments using
the HARV to drive the prototype Warrior and a high-speed surrogate UGV (a
modified, gas-powered R/C car used for testing) through a slalom course.
Stingray Video
The
Daredevil Project
|
|
|
|
|
|
I was the PI for the Daredevil Project, a recently-completed
Phase II SBIR project funded by TARDEC. For Daredevil, we developed the perception
techniques to allow robots to see through adverse weather (fog, rain, snow) and
sparse foliage. Robots often use LIDAR
(laser ranging) or vision to detect obstacles, but these sensors have difficulty
seeing through adverse weather and can be completely blocked by fog, smoke, or
dust. For Daredevil we used ultra
wideband (UWB) radar in combination with LIDAR and vision to allow the
Daredevil PackBot to avoid obstacles in all-weather conditions.
The images above show how UWB
radar can see through dense fog that blinds LIDAR and vision. On the left, the Daredevil PackBot is in
clear air, and both radar (green points) and LIDAR (red points) can see the
obstacles in the room. In these
conditions, LIDAR provides more precise range data with better angular
resolution. On the right, the Daredevil
PackBot is immersed in dense fog in the same room. The LIDAR is unable to penetrate the fog
beyond a depth of about one meter, but the UWB radar is completely unaffected. If the robot were only equipped with LIDAR and
vision, it would not be able to move safely in a fog-filled environment, but
using UWB radar, the Daredevil PackBot is able to successfully avoid obstacles
even in dense fog.
For more details, see my
papers for ICRA 2010 and SPIE Unmanned Systems 2010.
The Wayfarer Project
I was also the PI for the TARDEC-funded Wayfarer Project, a
two-year, $1.3 million effort to develop autonomous urban navigation
capabilities for man-portable mobile robots, such as the iRobot PackBot.
We equipped two Wayfarer PackBot prototypes with stereo vision and LIDAR to
perform autonomous reconnaissance missions in urban terrain, including
GPS-denied areas. The new ruggedized Wayfarer navigation payload is shown
above left, and a 3D map generated by the Instant Scene Modeler (iSM) during
perimeter reconnaissance is shown above right. (iSM was developed by Stephen Se at MDA Corporation and
was integrated with Wayfarer at iRobot
Corporation.)
Wayfarer Videos
- Rugged
Navigation Payload (3 MB) (high-quality
20 MB)
- Perimeter Reconnaissance with iSM 3D Mapping
- Route Reconnaissance
(7 MB) (high-quality 25 MB)
- Perimeter Reconnaissance
(16 MB) (high-quality 31MB)
including:
- Perimeter Following
- Mapping
- Obstacle Avoidance
- Automatic Flipper Deployment
- Indoor Mapping (6 MB) (high-quality11MB)
- Outdoor Obstacle Avoidance
(10 MB) (high-quality
20MB)
Previous Research
View the robots I've developed in my Robot Gallery.
I've previously conducted research and development in mobile robotics at:
- Naval Research Laboratory
- Institute for the Study of Learning
and Expertise
- Case
Western Reserve University
- NASA Kennedy Space Center
- Jet Propulsion Laboratory
- Hughes Research Laboratories
While at the Naval Research Laboratory, I developed frontier-based exploration, a technique that allows mobile robots to explore and map unknown environments.
Selected
Publications
All-Weather
Perception for Man-Portable Robots Using Ultra-Wideband Radar
Brian Yamauchi, Proceedings of the 2010 IEEE International Conference on
Robotics and Automation (ICRA 2010),
Anchorage, AK, May 2010
Fusing
Ultra-Wideband Radar and LIDAR for Small UGV Navigation in All-Weather
Conditions
Brian Yamauchi, Proceedings of SPIE Vol. 7692 (DS117): Unmanned Systems
Technology XII, Orlando, FL, April 2010
Daredevil: Ultra Wideband Radar Sensing for Small
UGVs
Brian Yamauchi, Proceedings of SPIE Vol. 6561: Unmanned Systems Technology
IX, Orlando, FL, April 2007
Autonomous
Urban Reconnaissance Using Man-Portable UGVs
Brian Yamauchi, Proceedings of SPIE Vol. 6230: Unmanned Systems Technology
VIII, Orlando, FL, April 2006
Wayfarer: An Autonomous Navigation
Payload for the PackBot
Brian Yamauchi, Proceedings
of AUVSI Unmanned Vehicles North America 2005, Baltimore, MD, June 2005
The
Wayfarer Modular Navigation Payload for Intelligent Robot Infrastructure
Brian Yamauchi, Proceedings of SPIE Vol. 5804: Unmanned Ground Vehicle
Technology VII, Orlando, FL, March 2005
Griffon:
A Man-Portable Hybrid UGV/UAV
Brian Yamauchi and Pavlo Rudakevych, Industrial Robot, Vol. 31, No. 5,
pp. 443-450, 2004
PackBot: A Versatile Platform for Military Robotics
Brian Yamauchi, Proceedings of SPIE Vol. 5422: Unmanned Ground Vehicle
Technology VI, Orlando, FL, April 2004
Integrating
Exploration and Localization for Mobile Robots
Brian Yamauchi, Alan Schultz, and William Adams, Adaptive Behavior, Vol.
7, No. 2, Spring 2000
Mobile Robot
Exploration and Map-Building with Continuous Localization
Brian Yamauchi, Alan Schultz, and William Adams, Proceedings of the 1998
IEEE International Conference on Robotics and Automation, Leuven, Belgium,
May 1998, pp. 3715-3720
Frontier-Based
Exploration Using Multiple Robots
Brian Yamauchi, Proceedings of the Second International Conference on
Autonomous Agents (Agents '98), Minneapolis, MN, May 1998, pp. 47-53
A
Frontier-Based Approach for Autonomous Exploration
Brian Yamauchi, Proceedings of the 1997 IEEE International Symposium on
Computational Intelligence in Robotics and Automation, Monterey, CA, July
1997, pp. 146-151
Place
Recognition in Dynamic Environments
Brian Yamauchi and Pat Langley, Journal of Robotic Systems, Special
Issue on Mobile Robots, Vol. 14, No. 2, February 1997, pp. 107-120
Spatial Learning
for Navigation in Dynamic Environments
Brian Yamauchi and Randall Beer, IEEE Transactions on Systems, Man, and
Cybernetics - Part B: Cybernetics, Special Issue on Learning Autonomous
Robots, Vol. 26, No. 3, June 1996, pp. 496-505
iRobot marks and media are used with permission. All rights reserved. Most content and media hosted on robotfrontier.com is available for public, private, and media use, on the condition that the content or media is used with appropriate attribution. Please contact me for permission to use any content or media from robotfrontier.com in your own work.